arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.01613cs.LGcs.AI
收藏

一种用于生产级大语言模型的实用张量网络压缩管道

A Practical Tensor-Network Compression Pipeline for Production-Scale Large Language Models

Minima AI, Inc.(Minima AI 公司)

Sergii Kozyrev, Davyd Maiboroda

AI总结:

Minima是一种用于大语言模型的实用张量网络压缩管道,通过结构压缩和自定义内核提升推理效率和吞吐量。

摘要:

大语言模型在部署中受到GPU内存和推理延迟的限制。我们提出了Minima,一种生产级压缩管道,它学习在何处以及如何对Transformer进行结构压缩,并将这种压缩转化为实际的服务增益。Minima训练了一个轻量级的卷积预测器来估计层级和补丁级的敏感度,然后对低敏感度区域应用 Tucker、张量列车和张量环分解的混合方法,进行一次短的修复微调,并使用自定义的Triton和CUDA内核执行所得到的运算符。减少的内存足迹使能够使用小型草稿模型和更大的验证器进行推测解码。在Qwen3-32B模型上,8k-token上下文窗口下,Minima将峰值VRAM从64GiB减少到40GiB。对于单个活跃请求,吞吐量从40个token每秒(基线)增加到50个token每秒(Minima)和75个token每秒(Minima加推测解码)。在50个并行请求下,吞吐量分别为34、44和53个token每秒,显示Minima在高并发情况下即使在推测解码增益压缩时仍有效。我们把Minima与最近的张量网络、低秩加量化和跨层共享方法进行比较,并认为它是一个朝着更激进的结构压缩迈出的实用步骤,通过共享张量骨干和微小的每层适配器。

英文摘要:

Large language models are limited in deployment by GPU memory and inference latency. We present Minima, a production compression pipeline that learns where and how to structurally compress a Transformer and turns that compression into real serving gains. Minima trains a lightweight convolutional predictor to estimate layer- and patch-level sensitivity, applies a mixture of Tucker, tensor-train, and tensor-ring decompositions to low-sensitivity regions, performs a short healing fine-tune, and executes the resulting operators with custom Triton and CUDA kernels. The reduced memory footprint enables speculative decoding with a small draft model and a larger verifier. On Qwen3-32B at an 8k-token context window, Minima reduces peak VRAM from 64 GiB to 40 GiB. For a single active request, throughput increases from 40 tokens per second (baseline) to 50 tokens per second (Minima) and 75 tokens per second (Minima with speculative decoding). Under 50 parallel requests, throughput is 34, 44, and 53 tokens per second respectively, showing that Minima remains effective under high concurrency even when speculative decoding gains compress. We position Minima relative to recent tensor-network, low-rank plus quantization, and cross-layer sharing methods, and argue that it is a practical step toward more aggressive structural compression via shared tensor backbones with tiny per-layer adapters.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑